Infrastructure-Free Indoor Pedestrian Tracking with Smartphone Acoustic-Based Enhancement
Total Page:16
File Type:pdf, Size:1020Kb
sensors Article Infrastructure-Free Indoor Pedestrian Tracking with Smartphone Acoustic-Based Enhancement Chao Liu *, Sining Jiang, Shuo Zhao and Zhongwen Guo Department of Information Science and Engineering, Ocean University of China, Qingdao 266100, China; [email protected] (S.J.); snickerfi[email protected] (S.Z.); [email protected] (Z.G.) * Correspondence: [email protected]; Tel.: +86-186-602-26768 Received: 16 April 2019; Accepted: 27 May 2019; Published: 29 May 2019 Abstract: Indoor pedestrian tracking has been identified as a key technology for indoor location-based services such as emergency locating, advertising, and gaming. However, existing smartphone-based approaches to pedestrian tracking in indoor environments have various limitations including a high cost of infrastructure constructing, labor-intensive fingerprint collection, and a vulnerability to moving obstacles. Moreover, our empirical study reveals that the accuracy of indoor locations estimated by a smartphone Inertial Measurement Unit (IMU) decreases severely when the pedestrian is arbitrarily wandering with an unstable speed. To improve the indoor tracking performance by enhancing the location estimation accuracy, we exploit smartphone-based acoustic techniques and propose an infrastructure-free indoor pedestrian tracking approach, called iIPT. The novelty of iIPT lies in the pedestrian speed reliability metric, which characterizes the reliability of the pedestrian speed provided by the smartphone IMU, and in a speed enhancing method, where we adjust a relatively less reliable pedestrian speed to the more reliable speed of a passing by “enhancer” based on the acoustic Doppler effect. iIPT thus changes the encountered pedestrians from an“obstacle” into an “enhancer.” Extensive real-world experiments in indoor scenarios have been conducted to verify the feasibility of realizing the acoustic Doppler effect between smartphones and to identify the applicable acoustic frequency range and transmission distance while reducing battery consumption. The experiment results demonstrate that iIPT can largely improve the tracking accuracy and decrease the average error compared with a conventional IMU-based method. Keywords: infrastructure-free indoor pedestrian tracking; acoustic signal; Doppler effect; Inertial Measurement Unit; arbitrary wandering 1. Introduction Considered as the last step of mobile service, indoor pedestrian tracking (IPT) has been widely utilized by researchers for the study of indoor location-based applications and services. For example, shopping mall navigation provides convenient shopping guidance and marketing services for customers. Hospitals leverage IPT systems to locate their specific patients in case of sudden accidents. Without IPT applications, people can get lost when they are trying to find their car in a large parking lot. Due to the building materials, GPS is unavailable for IPT, leading to a constant emerging of various smartphone-based tracking methods that have the capability of obtaining location information via Wi-Fi [1], Bluetooth [2,3], LED [4,5], and other techniques. Each smartphone-based IPT method has its pros and cons, and how to develop a low-cost, reliable, and accurate IPT system is still an outstanding research problem. Many attempts have been made in recent years. Yin et al. [6] utilized radio and visual features of a diagram of Wi-Fi fingerprints to realize IPT, but this still requires intensive labor for fingerprint collection. By leveraging additional beacons, Xiang et al. [7] combined Bluetooth and crowdsourcing to estimate pedestrian locations. Sensors 2019, 19, 2458; doi:10.3390/s19112458 www.mdpi.com/journal/sensors Sensors 2019, 19, 2458 2 of 16 In [8], a generic indoor localization framework is proposed based on existing lighting infrastructure, which is sensitive to moving obstacles. Limited by the demand for extra infrastructures, these methods suffer a relatively high cost. In order to realize infrastructure-free IPT, researchers leverage smartphone in-built sensors, such as inertial measuring units (IMUs), to estimate a pedestrian’s location. IMUs in commercial smartphones have been widely utilized for pedestrian dead-reckoning (PDR) [9], which contain multiple sensors such as accelerometers, gyroscopes, and sometimes magnetometers, and are capable of tracking the movement of pedestrians by detecting steps, estimating stride lengths, and the directions of motion. However, the tracking accuracy of PDR decreases severely when the pedestrian walks around arbitrarily, causing accumulative error. To improve PDR accuracy, researchers [10,11] focus on the enhancement of a pedestrian’s steps detection and stride length estimation, but they still cannot satisfy the actual accuracy requirement due to the complex indoor environment and the pedestrian’s arbitrary walking. Recently, acoustic-based IPT [12] has shown its unique advantages of accurate positioning and direction finding. To be more specific, ultrasonic-based IPT systems [13,14] perform well at estimating location with a high degree of accuracy. Ultrasonic frequency is normally 20–44 kHz (inaudible for human) and easily removes noise. Due to the low velocity of ultrasonic wave traveling in air, high accuracy time of flight (ToF) measurement is allowed. Moreover, the Doppler effect has been widely utilized to track mobile objects [15,16]. However, there are two main limitations that affect the performance of ultrasonic-based IPT. Firstly, similar to other smartphone-based IPT methods, extra infrastructures and self-designed devices are required to guarantee the tracking accuracy by emitting and receiving ultrasonic signals, which is time-consuming and has a high cost. Secondly, ultrasonic signals can be easily blocked by mobile obstacles due to its physical features, causing reflection and multi-path problems, ultimately decreasing tracking accuracy. According to [17], smartphones can emit and receive ultrasonic signals between each other. What if we utilized smartphones only to accomplish IPT by combining the advantages of ultrasonic and IMU methods? In this paper, we propose iIPT, an infrastructure-free pedestrian tracking approach, by combining smartphone-based acoustic and IMU techniques. A metric called pedestrian speed reliability is presented to characterize the reliability of the pedestrian speed provided by the smartphone IMU. We adjust a relatively less reliable pedestrian speed to a more reliable speed of a passing by ”enhancer“ based on the acoustic Doppler effect, thus changing the encountered pedestrians from an “obstacle“ into an ”enhancer“. Extensive real-world experiments in indoor scenarios have been conducted to verify the feasibility of realizing the acoustic Doppler effect between smartphones and to identify the applicable acoustic frequency range and transmission distance while reducing battery consumption. The experiment results demonstrate that iIPT can largely improve the tracking accuracy and decrease the average error compared with a conventional IMU-based method. We have made several contributions as follows: • We present an infrastructure-free indoor pedestrian tracking approach by combining both smartphone-based acoustic and IMU techniques. • We present a pedestrian speed reliability metric that characterizes the reliability of the real-time pedestrian speed provided by smartphone IMU and reflects the arbitrariness of the pedestrian walking pattern. • We determine the capability and precision of measuring speed using sonic Doppler from a smartphone by designing a robot car whose speed could be controlled to compare the real speed. • We leverage the acoustic Doppler effect to adjust the relatively less reliable pedestrian speed to a more reliable speed of a passing by “enhancer” measured by IMU. • We implement comprehensive experiments to identify the applicable acoustic frequency range, transmission distance, and battery consumption and demonstrate that iIPT can largely improve the tracking accuracy and decrease the average error, compared with PDR. The remainder of this paper is organized as follows: Section2 illustrates our motivation and challenges for proposing iIPT. In Section3, an overview of iIPT is presented and Section4 elaborates Sensors 2019, 19, 2458 3 of 16 our indoor pedestrian tracking approach. Section5 evaluates the performance of iIPT in real situations. In Section6, we discuss about the limitations and unsolved problems for iIPT. Finally, Section7 provides a brief conclusion. 2. Motivation and Challenges In a common indoor environment, a pedestrian’s attention can be easily drawn by many places of interest, leading to different walking patterns. For indoor pedestrian tracking with smartphone IMUs, a sudden stop or speed up will increase the complexity of walking detection and lower the tracking accuracy. In this scenario, we present two important observations. Observation 1. A pedestrian usually walks arbitrarily with an unstable speed in an indoor environment. According to research [18,19] on human mobility models in indoor environments, pedestrian mobility can be influenced by “social forces,” including certain motion requirements, other, disturbing pedestrians, and various attractive effects. For example, customers are commonly attracted by clothing shops or restaurants when shopping in a mall. Visitors occasionally concentrate on gorgeous art work at exhibitions when they have great interest in it. These kinds of cases happen quite often in indoor environments and cause problems for IMU-based indoor pedestrian